Multi-objective Reinforcement Learning

Abstract

In this talk we present PQ-learning, a new Reinforcement Learning (RL) algorithm that determines the rational behaviours of an agent in multi-objective domains